Collaborative wireless mesh network architecture for internet of things in perspectives on reliable networking and mobility management

Abstract

In the near future, more and more devices will be capable of sending and receiving data automatically via the Internet. The vision of Internet of Things (IoT) is to embrace trillions of everyday physical objects that surround us as service providers and consumers by transforming them into an ecosystem of information that will enrich our lives. From small light bulbs to houses, the IoT is bringing more and more various things into the globally inter-connected digital world every day. While the IoT represents the convergence of advances in miniaturization, Internet connectivity, increased data storage capacity, and batteries, the realization of the IoT would not be possible without sensors / actuators and lightweight wireless connectivity. Since recent attempts to speed up the IoT research and facilitate the new wave of information technology have spurred the need for lightweight IP networking technologies for the resource-constrained physical things, 6LoWPAN standards attracts attention as one of such upcoming technologies which breathes life into everyday objects by enabling IP over resource-limited, low-power, and low-rate wireless networks. However, there are still barriers to provide "Smart" IoT services in our everyday life. Firstly, a reliable and seamless network connectivity of wide prevalence of smart things with high perceived values is indispensable prerequisite for high satisfaction of user experience. Secondly, since the IoT is a cybernetic representation of dynamic and mobile physical world, seamless mobile connectivity for the physical objects is an important challenge. To address them, this thesis presents lightweight wireless mesh network protocols for reliable communications with mobility in the IoT. As the cornerstone to enable IPv6-based wireless networking for the IoT devices, this dissertation is comprised of three parts: reliable communication which includes load balancing and data caching, mobility management, and IoT device platform for easy-construction & deployment. From the reliable communication aspects, the first part of this dissertation addresses a dynamic and distributed load balancing scheme in multi-gateway based 6LoWPAN. In a LoWPAN, a 6LoWPAN GW (GW) is a border router to bridge a LoWPAN to the Internet. Hence, all the network traffic should pass through the GW, so a GW as the traffic center may become the congestion-concentrated point. The network throughput and capacity are also limited by the GW itself. Even though we apply multiple GWs to remove the limitation, the imbalance of traffic load among GWs may restrict overall network performance without the expected enhancement. The load balancing between GWs is a key technology to reduce the traffic congestion thereby enlarging the network capacity, satisfying the capacity fairness among 6LoWPAN nodes, and improving network reliability. In this respect, this thesis proposes a dynamic and distributed load balancing scheme to achieve a global load fairness, motivated by water flow behavior, named MLEq (Multi-GW Load Balancing Scheme for Equilibrium). We implement our load balancing scheme in network simulator (ns-2) and it is verified by extensive simulations. The results show that MLEq improves network throughput and fairly distributes traffic load among GWs with minimizing network overhead. More importantly, we observed that MLEq increases network capacity and throughput linearly with the number of gateways. Secondly, this dissertation addresses a benefit-based distributed data caching algorithm for 6LoWPAN. In general, information access is very important in 6LoWPAN, where being equipped with sensors and providing data objects as a service, since the ultimate goal of 6LoWPAN is to provide information access to users. However, due to the low-rate wireless multi-hop nature of 6LoWPAN, service availability is limited by the link status and traffic load over multi-hop. In addition, low-power networking like duty-cycling mechanism may increase data access delay. Caching and replication of popular data objects can significantly improve the efficiency of information access by contributing the reduction of overall access time and network bandwidth usage. For this, this thesis proposes a benefit-based cache distribution algorithm (BCDA). The purpose of this work is to provide an efficient data caching algorithm which suggests the proper cache places, where maximizes the benefit in terms of data access cost and cache management cost. To justify the effectiveness, we implement our BCDA in ns-2 and show that the BCDA significantly outperforms other strategies in terms of access delay, service success ratio, signaling overhead, and memory overhead. From the mobility management aspects, this dissertation addresses a fast and seamless mobility management protocol for 6LoWPAN. As mobile devices come into wide use in the Internet, the mobile data traffic has grown rapidly and the demand for mobile services has increased globally over the past few years. The mobility management becomes one of the most important network services in the IoT. Since thing`s mobility behaviour directly inherits characteristics of portable devices (i.e., tiny, battery-powered, and wireless), the mobility management should be supported in a lightweight fashion. To address this, this dissertation proposes a new fast and seamless mobility management protocol for 6LoWPAN, named MARIO. To achieve this, MARIO introduces three important components, which are an adaptive polling based movement detection, a fast rejoin scheme for handoff management, and a multi-hop pointer forwarding scheme for location management. We develop an analytical model to evaluate the performance of MARIO and its signaling overhead. We then implement MARIO on top of our 6LoWPAN platform and conduct real-world experiments to validate its performance. The results highlight that the MARIO reduces signaling overhead and handoff delay and minimizes packet losses during handoffs significantly. Lastly, one important challenge in the IoT is how to achieve real-world deployment of 6LoWPAN to be a part of the IoT on the fly and on the cheap. Since 6LoWPAN requires additional efforts to connect a LoWPAN to the Internet through either wired or wireless medium, this results in expensive partial reconstruction of existing infrastructure and making its deployment difficult in real-world. However, until now how to achieve real-world deployment and its integration with the Internet on the fly and on the cheap is still foggy. Fortunately, we have witnessed successful deployment of WiFi APs anywhere and anytime as a part of Internet Infrastructure. Motivated by the successful deployment of WiFi APs, this thesis introduces two types of 6LoWPAN gateways to achieve real-world deployment of IoT devices with easy-construction and cheap-deployment.